Description: The Center has embarked on a systematic study of the changing mental health system. This core supports that effort along several dimensions through four projects that focus on the "meso" and direct service level of mental health services delivery. These levels address the implementation of disease management programs, the organization and management of patient care teams, and the rules for triage and eligibility for specialized services. This focus is seen as necessary, according to the researchers, due to the challenge of adapting the results of efficacy studies to real conditions under which mental health services are provided, and to improve the effectiveness of managed care strategies, and incentives. The studies follow a trajectory of methodologies that began with descriptive studies, followed by multivariate prospective studies to more complex randomized controlled trials. The subject of the proposed work is patient compliance, specifically medication compliance in different service settings. They propose to examine strengthening client/family interactions through building bonds of reciprocity; tracking portions of the mental health delivery systems through collaboration on the Health Care for Communities (HCC) survey and through state surveys of residential facilities. For Project 1A (p 306), it is proposed that compliance behavior be controlled for when evaluating experimental-control group differences in clinical and social outcomes at 6 and 12 weeks. It is presumed that compliance here means compliance with the compliance therapy. In any case, the investigators should be more specific about how they propose to adjust for compliance behavior given the pitfalls of adjusting for observed compliance when analyzing differences between randomized groups. Given the investigators? knowledge of instrumental variables, it is likely that they understand the use of this approach to adjust for non-compliance with the compliance therapy. Also for Project 1A (p 306), a cost analyses is proposed without specifics. The analysis plans for Projects 2 and 3 (p. 308-309, 311-312) are too general, especially given that analyses are planned in the first year of the grant (p. 317). The pilot data analysis proposed for Project 4 (page 315-316) appears to run the risk of possible non-compliance problems with the INCube program. An instrumental variable analysis may be appropriate here.